Carbon calabashes in a shock-produced carbon melt

Article abstract of DOI:10.1016/S0009-2614(03)00711-5

Carbon calabashes similar to those reported during decomposition of CH₄ between 900 and 1050 °C also formed after thermal decomposition of dolomite, CaMg(CO₃)₂ in a low-angle impact experiment that produced a carbon melt and complex gas phase between 2700 and 3700 °C. The interrelationships among various carbon forms in this experiment support that the calabashes are quenched carbon-melt bubbles at high gas to liquid ratios. Carbyne forms of carbon that crystallized within the quenched-liquid shell of the bubbles did not require a spiral growth mechanism. Carbon calabashes are an oddity but not a new form of elemental carbon.

Full text of DOI:10.1016/S0009-2614(03)00711-5

Chemical Physics Letters 374 (2003) 464–470
www.elsevier.com/locate/cplett
Carbon calabashes in a shock-produced carbon melt
a,
Frans J.M. Rietmeijer , Peter H. Schultz , Theodore E. Bunch
b
c
*
a
Department of Earth and Planetary Sciences, University of New Mexico, Albuquerque, NM 87131-1116, USA
b
Department of Geological Sciences, Brown University, Providence, RI 02912, USA
Department of Geology and Bilby Research Center, Northern Arizona University, Flagstaff, AZ 86011, USA
c
Received 18 March 2003
Abstract
Carbon calabashes similar to those reported during decomposition of CH
4
between 900 and 1050 °C also formed
after thermal decomposition of dolomite, CaMg(CO in a low-angle impact experiment that produced a carbon melt
3 2
)
and complex gas phase between 2700 and 3700 °C. The interrelationships among various carbon forms in this ex-
periment support that the calabashes are quenched carbon-melt bubbles at high gas to liquid ratios. Carbyne forms of
carbon that crystallized within the quenched-liquid shell of the bubbles did not require a spiral growth mechanism.
Carbon calabashes are an oddity but not a new form of elemental carbon.
Ó 2003 Elsevier Science B.V. All rights reserved.
It has been said before. Carbon is a very ver-
satile element that responds to kinetically con-
trolled conditions with the formation of a
multitude of structures such as ÔgrapheneÕ sheets of
hexagonal rings [1], carbon onions [2], and open
and closed nanotubes [3]. The substitution of
hexagons in aromatic carbon rings by pentagons
leads to curving and closed carbon structures such
as the C60 fullerene molecule [4]. Decomposition of
At temperatures between 900 and 1050 °C this
process produced (1) filled spheres wherein the
graphitic carbon core is disconnected from its
uniform shell ꢀ10 nm thick, and (2) (rare) gra-
phitic hollow carbon calabashes that can be both
spherical and dumbbell-shaped [6]. The calabashes
are hollow, closed graphitic shells ꢀ0.8 lm in di-
ameter and occur as individual structures that
contain trapped gases that could be oxygen, hy-
drogen and/or methane [6]. The massive spheres
and hollow calabashes are not in contact nor en-
CH
4
at 1100 °C using the mixed-valent oxide-cat-
alytic carbonization process yielded massive gra-
phitic spheres of uniform diameter wherein wavy
proto-crystallites arise from combinations of
hexagonal, pentagonal and heptagonal rings [5].
2
capsulate the MnO catalyst and they could be a
new carbon structure, albeit that a spiral growth
model could not account for their formation [6].
This Letter reports the finding of similar
spherical calabashes produced in impact experi-
ments designed to assess ejecta and vapor plume
production and evolution during major impact
*
Corresponding author. Fax: +1-505-277-8843.
E-mail address: fransjmr@unm.edu (F.J.M. Rietmeijer).
0
009-2614/03/$ - see front matter Ó 2003 Elsevier Science B.V. All rights reserved.
doi:10.1016/S0009-2614(03)00711-5

F.J.M. Rietmeijer et al. / Chemical Physics Letters 374 (2003) 464–470
465
events on the EarthÕs surface. One such impact,
which led to the extinction of the dinosaurs pene-
phases albeit somewhat ordered in physical struc-
tures, e.g., jets and clouds, containing multiple
chemical species. At these low angles, the peak
shock pressure in both target and projectile is
significantly reduced even though the degree of
thermal decomposition appears to increase [13].
This observation has led to the interpretation that
oblique impacts induce significant thermal de-
composition of the target material through high-
strain rate, frictional shear, heating and shock
heating. Such thermal decomposition processes
were also observed from experiments using a fric-
tion-melting apparatus, that is, even without shock
[11]. Under these conditions the reaction products
appear as frothy, melted materials near the crater
rather than as debris that was ejected at high
velocities.
3 2
trated a sedimentary dolomite, CaMg(CO ) , de-
posit at a site known as the Chicxulub crater.
The shock-produced debris included a mixture of
comminuted fragments with solid-state transfor-
mations plus quenched-melt dust and vapor con-
densed dust, as well as carbon phases resulting
from carbonate dissociation. The multiplicity of
co-occurring carbon phases, such as soot, poorly
graphitized carbon, graphite, diamond and fulle-
renes [7,8], or lonsdaleite and chaoite under oxi-
dizing conditions [9], reflects the dynamic chemical
and physical conditions during this short-lived
event and subsequent heat dissipation.
The experiments fired a 0.375-g Al sphere at
4
.9 km/s into a sample of natural dolomite at a 15°
angle from the horizontal using the NASA Ames
Vertical Gun Range [10,11]. The impact-generated
hypervelocity jets and plumes were monitored for
The recovered dolomite fragments were coated
by a thin, sponge-like black layer ꢀ50 lm thick
[14] with information on the fundamental chemical
and physical processes operating as a function of
time and location in the evolving experiment.
Ultrathin (80–100 nm) sections of debris were
prepared for characterization using a JEOL
2000FX analytical electron microscope (AEM)
equipped with an energy dispersive spectrometer
for detection of elements with Z > 11. The solids
were identified using chemical composition and
crystallographic properties determined by selected
area electron diffraction and high-resolution
transmission electron microscopy (TEM). The
AEM plus Auger and Raman spectroscopy [14]
data show angular dolomite fragments. Larger
fragments up to 500 ꢁ 160 nm show structural
disorder (subgrains), thermally induced erosion
and melting. Dolomite decomposition, viz.
2
0 ls by impact flash spectroscopy (IFS) and high-
speed imaging. The IFS spectrum showed a very
high continuum background, which is most likely
caused by incandescent melt droplets and solid
fragments from the projectile and dolomite target.
It can be fitted with a single-temperature black
body curve between 2700 and 3700 °C [10] con-
ducive to carbonate vaporization. The spectral line
emissions also indicate the presence of CaO, MgO
and Na in the gas that probably also contained CO
[
10]. In this low-angle impact experiment the vapor
may have been mixed with or masked by frag-
mented material and melt droplets in the jets
[
10,12]. This type of experiments will also generate
an expanding vapor cloud that evolves differently
and separately from the hypervelocity jets. It is a
dominant feature at impact angles below 30° [13].
The internal energy of this ÔcoolÕ vapor is just
above the carbonate thermal decomposition tem-
perature. The vapor composition after ꢀ150 ls
was determined using a photographic diffraction
spectrograph, viz. AlO (projectile), neutral species
CaMg(CO
3
)
2
¼ CaO + MgO + 2CO
2
, lead to dense
pockets of either ultrafine-grained (<10 nm) ox-
ides (comminution) or coarse-grained oxides
ranging from ꢀ10 to 115 nm (mean: 31 nm). Those
smaller than ꢀ50 nm are rounded; larger grains
are typically euhedral crystals. Aerodynamic
shapes of CaO and MgO grains, e.g., elongated
teardrops, suggest they are quenched-liquid drop-
lets (comminution and melting). The size distri-
bution shows that the euhedral grains grew via
Ostwald ripening of smaller entities. The elemental
carbon forms included:
(
CaI; NaI, an unidentified contaminant) and the
carbonate dissociation products CaO and CO [13];
in the dolomite experiment MgI and MgO also
continued to generate spectral emission lines to
later times. In this highly dynamic shock experi-
ment matter occurred in the solid, liquid and gas

466
F.J.M. Rietmeijer et al. / Chemical Physics Letters 374 (2003) 464–470
(
(
1) Carbon onions (extremely rare).
2) Abundant amorphous carbon balls with a
small CaO or MgO core.
3) Graphitic rims on the coarse-grained MgO
(5) Abundant hollow (TEM contrast analysis)
closed structures with a continuous graphitic
shell.
(
(a) Common spheres of 28 to 170 nm in di-
ameter (mean ¼ 69 nm; 1r ¼ 31.5 nm). The
shell thickness ranges from 2.7 to 26.2 nm
(mean ¼ 6.4 nm). The ratio of the shell width
to the radius of the structure ranges from
0.125 to 0.31 (mean ¼ 0.19 nm; 1r ¼ 0.05
nm). In a few spheres this ratio is 0.5. In
the ultrathin sections they appear as ring
structures with a very thick wall (Fig. 2).
(b) ÔSoap bubbleÕ structures that are 45 up to
308 nm in diameter (mean ¼ 114 nm;
1r ¼ 67.5 nm) (Fig. 3). The shell thickness
ranges from 2.3 to 6.5 nm. The shell width
and CaO grains. These rimmed grains resemble
the Ni-particles with graphitic rims that cata-
lyzed the transformation of amorphous carbon
to graphite between 400 and ꢀ725 °C in the
Ni–C system [15] and using a Ni/MgO catalyst
[
16]. The graphitic rim in several rimmed grains
is separated from the trailing end of the aerody-
namically shaped core (Figs. 1 and 2).
(
4) (Rare) massive, amorphous single carbon
balls up to 60 nm in diameter, and a few carbon
dumbbells, with a continuous graphitic rim (up
to 20 nm thick) (Fig. 1).
Fig. 1. Transmission electron micrograph of an ultrathin section showing condensed carbon-rich entities after shock-induced disso-
ciation of dolomite. Mostly at the top are several entities with a dark core of MgO and CaO surrounded by a graphitic carbon rim. The
sub-spherical entity on the right-hand side has an amorphous carbon core surrounded by a graphitic rim of concentric fringes
highlighted by radial strain contrast features. Both objects to the left are hollow spheres with a broad graphitic carbon rim. The
seemingly dark core of the central entity is due to an underlying metal-oxide/carbon entity.

F.J.M. Rietmeijer et al. / Chemical Physics Letters 374 (2003) 464–470
467
Fig. 2. Transmission electron micrograph showing two large spheres, and a small, hollow spheres with a graphitic carbon rim of
variable thickness of concentric fringes in an ultrathin section. Also shown several entities with a massive (dark) metal-oxide core and a
graphic carbon rim. The objects are off the proper orientation in this image to show the graphic fringes.
to radius ratio ranges from 0.03 to 0.12;
the function, y (ratio) ¼ )0.048 ln (x; diame-
lated in some plausible manner. The shell width on
the closed structures shows no correlation with
either the closed shell diameter or the shell width
to radius ratio, which is consistent with the dy-
namic nature of the impact process including the
dissipation of thermal energy.
The lattice fringes in the shells of the hollow,
closed spheres are typically well defined in a con-
centric circular pattern. In the calabashes they can
have a wavy appearance due to bubble expansion.
In the former the lattice fringes range from 1.0 to
1.5 nm (mean ¼ 1.3 nm; 1r ¼ 0.2 nm). Those in the
shell of the calabashes range from 0.9 to 1.5 nm
(mean ¼ 1.1 nm; 1r ¼ 0.2 nm). Statistically there is
no evidence that the two sample groups belong to
populations having different means. That is, the
average lattice fringe spacing is 1.2 Æ 0.2 nm but
the observed range is larger than the error of
2
ter) + 0.302 (R ¼ 0:70), is consistent with
the thinning of a (liquid) shell as a function
increasing bubble size. The bubbles are only
a very small fraction of the hollow closed car-
bons. They occur mostly isolated but adjoin-
ing neighbors in clusters can be distorted.
The largest bubbles may have elongated
shapes or form collapsed bubble structures.
In all fundamental properties that we deter-
mined, the second type of hollow, closed carbon
structures with a continuous graphitic shell re-
semble the hollow calabashes of unknown origin
reported by [6]. We find that they co-occur with a
more abundant hollow, closed carbon structure,
which is not fundamentally different from the
calabashes to which they must be genetically re-

468
F.J.M. Rietmeijer et al. / Chemical Physics Letters 374 (2003) 464–470
Fig. 3. Transmission electron micrograph showing a domain of high density of hollow carbon calabashes in this ultrathin section
surrounded a area of high density metal-oxide/carbon spheres (upper right-hand portion) and massive carbon deposits (top-left and
lower right-hand corner). The embedded large (sub-)equant grains mostly in the left side of the image are condensed MgO and CaO
grains.
measurement (ꢀ15% rel.). One possible interpre-
crystalline carbon [18]. Metastable carbynes are
supported by the fact that carbynes are not ex-
perimentally verifiable equilibrium carbon forms
[19,20]. There is a possibility that the carbynes in
the impact experiment contain some amount of
oxygen [18,21].
tation is that these fringes represent the carbyne
(
0 0 1) lattice plane. Carbynes are triple- and dou-
ble-bonded linear carbon polytypes parallel to the
crystallographic c-axis that were reported in high
temperature treatment of graphite and the hkl
(
0 0 1) spacing of the majority of these carbons
The shells show radial strain contrast perpen-
dicular to the crystalline carbon basal planes that
look identical to the contrast zones in graphitic
rims due to the catalyzed graphitization of amor-
phous carbon on Ni-grains [16,17]. This kind of
contrast is in fact common to all bent layered
mineral structures such as layer silicates. In the
crystalline shells this dislocation strain contrast
range from 0.96 to 1.54 nm [16]. The values indi-
cate either a chaotic melange of carbynes (includ-
ing chaoite) or only preferred carbynes, e.g.,
carbon-VI and carbon-IX, as indicated by the
average value. This interpretation is consistent
with highly kinetic conditions in the impact ex-
periment that favor the formation of metastable

F.J.M. Rietmeijer et al. / Chemical Physics Letters 374 (2003) 464–470
469
developed in the growing basal lattice fringes
along the circular perimeter of the hollow struc-
tures, including calabashes. It is possible that dis-
locations resulted from the misfit between basal
layers in crystallizing shells [22]. It is unclear why it
would cause radial contrast features with variable
spacing within a shell. In other calabashes high
lattice strain in curving shells was released by
the formation of kink-bands creating partially
polygonal shell segments. The strain contrast
indicates that the closed shells did not form by
substituting pentagons in growing structures of
hexagons because this process results in strain-free
curvature [23–25].
liquid–carbon to gas ratios and lateral growth
of graphene layers within the quenched-liquid
shell where crystallographic ordering produced
metastable crystalline carbons, i.e., carbynes.
Calabashes are the by-products of kinetically
controlled elemental carbon-forming processes
when a gas phase percolated through a carbon
melt. The calabashes did not form by a spiral
growth mechanism [6] and do not represent a new
form of elemental carbon. Impact-generated car-
bonate melts are known to occur in terrestrial
crater deposits [30,31]. Calabashes in these and
other impact crater deposits may provide for ad-
ditional indicators of the impact shock regime.
The state of carbon in the shock experiment
holds the key to the origin of calabashes. Carbon
resulted from CO
shock regime as CO
M + CO
2
dissociation in the initial
¼ C + O , or a reaction,
) 2MOðgasÞ + C, catalyzed by neutral
Acknowledgements
2
2
2
2
The analytical work was performed in the
Electron Microbeam Analyses Facility, Depart-
ment of Earth and Planetary Sciences (UNM),
where Fleur Rietmeijer-Engelsman provided pho-
tographic support. The hypervelocity experiments
were performed at the NASA Ames Vertical Gun
Range where we gratefully acknowledge the tech-
nical support. This work was supported by Grants
NAG5-11762 (FJMR), NAG5-11538 (PHS) and
344-38-12-02 (TEB) from the National Aeronau-
tics and Space Administration.
Mg and Ca gas species, or both. The thermal re-
gimes in the jets and carbon and oxygen-bearing
vapor cloud [12,13] straddle the metastable solid–
liquid phase boundaries in the carbon phase
diagram [20, Fig. 10], thereby supporting the
presence of a metastable liquid carbon phase
during the experiment in addition to a sizeable gas
phase.
Hollow spheres of Fe-oxide occur on meteorite
fusion crusts as a result of melt–vapor interactions
during deceleration in the EarthÕs atmosphere [26]
and in lightning-strike induced evaporation and
condensation experiments [27]. Hollow alumino-
silica spheres are common among industrial coal-
fired fly ash [28] while hollow microspherules occur
in some active volcanoes [29]. These hollow
spheres have in common the fact that they are
invariably from highly dynamic environments with
fast interactions between variable amounts of
liquids and gases. In the impact experiment, an
oxygen-rich gas dynamically mixed in the carbon
melt was rapidly released. As a result, bubbles
formed with a thin, liquid carbon shell that rapidly
quenched into a solid, amorphous carbon shell.
The subsequent, kinetically controlled devitrifica-
tion was driven by the free energy difference
between the amorphous and crystalline carbons.
The formation of the hollow closed structures in-
cluding the calabashes required rapidly changing
References
[1] T.J. Bernatowicz, R. Cowsik, P.C. Gibbons, K. Lodders,
B. Fegley Jr., S. Amari, R.S. Lewis, Astrophys. J. 472
(
1996) 760.
[
[
2] P.P.K. Smith, P.R. Buseck, Science 212 (1981) 322.
3] S. Iijima, T. Ichihashi, Y. Ando, Nature 356 (1992) 776.
[4] H.W. Kroto, K. McKay, Nature 331 (1988) 328.
[5] Z.L. Wang, Z.C. Kang, J. Phys. Chem. 100 (1996)
1
7725.
[
[
6] Z.L. Wang, J.S. Yin, Chem. Phys. Lett. 289 (1998) 189.
7] T.E. Bunch et al., Geol. Soc. Am. Spec. Paper 339 (1999)
3
23.
[8] A. Rotundi, F.J.M. Rietmeijer, L. Colangeli, M. Mennella,
P. Palumbo, E. Bussoletti, Astron. Astrophys. 329 (1998)
1
087.
9] Y. Miura, M. Okamoto, Antarctic Meteorites XXI (1996)
07.
[
1
[10] S. Sugita, P.H. Schultz, M.A. Adams, J. Geophys. Res. 103
(E8) (1998) 19427.

470
F.J.M. Rietmeijer et al. / Chemical Physics Letters 374 (2003) 464–470
[
[
11] C. Van der Bogert, P.H. Schultz, J. Spray, Meteoritics
Planet. Sci. (abstract) (1999) A115.
12] S. Sugita, P.H. Schultz, J. Geophys. Res. 104 (E12) (1999)
[23] C. Clinard, J.N. Rouzaud, S. Delpeux, F. Beguin, J.
Conard, J. Phys. Chem. Solids 55 (1994) 651.
[24] W.A. de Heer, D. Ugarte, Chem. Phys. Lett. 207 (1993)
480.
30825.
[
[
13] P.H. Schultz, J. Geophys. Res. 101 (E4) (1996) 21117.
14] T.E. Bunch, L. Becker, P.H. Schultz, W.S. Wolbach, Lunar
Planetary Instit. Contrib. 955 (1997) 5.
[25] J.B. Howard, K.D. Chowdhury, J.N. Vander Sande,
Nature 370 (1994) 603.
[26] F.J.M. Rietmeijer, I.D.R. Mackinnon, J. Geophys. Res. 87
(Suppl.) (1984) B597.
[
[
[
15] R. Lamber, N. Jaeger, G. Schulz-Ekloff, Surf. Sci. 197
(
1988) 402.
16] E. Tracz, R. Scholz, T. Borowiecki, Appl. Catalysis 66
1990) 133.
17] R.N. Heimann, J. Kleiman, N.M. Salansky, Nature 306
1983) 164.
[27] F.J.M. Rietmeijer, J.M. Karner, J.A. Nuth III, P.J.
Wasilewski, Eur. J. Mineral. 11 (1999) 181.
[28] R.T. Hemmings, E.E. Berry, Mat. Res. Soc. Symp. Proc.
113 (1988) 3.
(
(
[29] R. Lef ꢀe vre, A. Guadichet, M.A. Billon-Galland, Nature
322 (1986) 817.
[
[
[
[
[
18] F.J.M. Rietmeijer, Meteoritics 28 (1993) 242.
19] P.P.K. Smith, P.R. Buseck, Science 216 (1982) 984.
20] F.P. Bundy, J. Geophys. Res. 85 (B12) (1980) 6930.
21] F.J.M. Rietmeijer, Carbon 29 (1991) 669.
22] P.J.F. Harris, M.L.H. Green, S.C. Tsang, J. Chem. Soc.,
Faraday Trans. 89 (1993) 1189.
[30] T. Bunch, R.A.F. Grieve, P. Lee, C.P. McKay, J.W. Rice.
J.W. Schutt, A. Zent, Lunar Planet Sci. Conf. XXIX
(abstract) (1998) #1307, Lunar and Planetary Institute,
Houston, TX, CD-ROM.
[31] G.R. Osinski, Earth Planet. Sci. Lett. 194 (2001) 17.